Testing innovative crowd analysis to improve public safety in crowded places. This project uses advanced analytics technologies at Sydney Olympic Park.
Introduction
This project uses advanced analytics technologies at Sydney Olympic Park. The aim is to improve crowd safety during big events.
The trial will use de-identified footage taken from existing security cameras. Information from this footage will be combined with other data, including environment data and social media.
This will give the project team real-time updates about any safety or security risks. It will help to make the experience better for people attending events in the area.
This trial makes sure personally identifiable information remains protected. The team uses privacy-enhancing safety measures and controls to make sure nobody can be identified. The NSW Government has carefully checked this project for ethical and privacy concerns.
This project is delivered at the same time as the Digital Trust in Places and Routines pilot.
Challenge
Sydney Olympic Park is a major events and sporting precinct with large urban parkland. It is a fast-growing neighbourhood, with over 20,000 daily residents, workers and students. The Park attracts over 10 million visitors and stages 5000 events each year, often with multiple events occurring at the same time.
Public safety and protection of crowded places is a major risk for the NSW Government and Sydney Olympic Park Authority (SOPA), particularly during major events.
SOPA has a highly experienced team. However, there is a heavy reliance on the making of ‘judgement calls’ based on their experience and observations to manage crowded places. New technologies, including crowd modelling systems, have identified opportunities to use data to inform proactive decision making for public safety.
Solution
Creating Safer Crowded Places is a complex, multi-modality sentiment score regression model. It uses the latest methods in statistical machine learning to improve decision making around public safety.
The trial applies privacy-enhancing safeguards and controls to ensure that individuals are never identifiable.
This new technology will build on crowd modelling systems that draw data from SOPA’s existing CCTV network. It combines this with additional data from other sources, including social media and environmental systems. The platform then alerts the SOPA operational team to changes in customer sentiment about places. This allows SOPA to focus on proactive crowd-management techniques.
Expected Outcomes
Smart Places outcome areas:
- safety and security
- collaboration and connection.
This project will also:
- improve safety ratings at Sydney Olympic Park, and a corresponding decrease in safety incidents at large-scale events
- deliver operational efficiencies and ultimately, an improved experience for residents and visitors.
